COSTS AND EFFECTS OF ARKANSAS' LONG-TERM CARE DEMONSTRATION PROJECT (MEDICAID, CHANELLING, COMMUNITY ASSESSMENT)

The purpose of this study was to assess the impact of the Arkansas Long-Term Care Demonstration Project upon Arkansas' Medicaid expenditures and upon the clients it serves. A Retrospective Medicaid expenditure study component used analyses of variance techniques to test for the Project's effects upon aggregated expenditures for 28 demonstration and control counties representing 25 percent of the State's population over four years, 1979-1982.^ A second approach to the study question utilized a 1982 prospective sample of 458 demonstration and control clients from the same 28 counties. The disability level or need for care of each patient was established a priori. The extent to which an individual's variation in Medicaid utilization and costs was explained by patient need, presence or absence of the channeling project's placement decision or some other patient characteristic was examined by multiple regression analysis. Long-term and acute care Medicaid, Medicare, third party, self-pay and the grand total of all Medicaid claims were analyzed for project effects and explanatory relationships.^ The main project effect was to increase personal care costs without reducing nursing home or acute care costs (Prospective Study). Expansion of clients appeared to occur in personal care (Prospective Study) and minimum care nursing home (Retrospective Study) for the project areas. Cost-shifting between Medicaid and Medicare in the project areas and two different patterns of utilization in the North and South projects tended to offset each other such that no differences in total costs between the project areas and demonstration areas occurred. The project was significant ((beta) = .22, p < .001) only for personal care costs. The explanatory power of this personal care regression model (R('2) = .36) was comparable to other reported health services utilization models. Other variables (Medicare buy-in, level of disability, Social Security Supplemental Income (SSI), net monthly income, North/South areas and age) explained more variation in the other twelve cost regression models. ^

DISCHARGE PLANNING: A PROCESS MODEL

As the requirements for health care hospitalization have become more demanding, so has the discharge planning process become a more important part of the health services system. A thorough understanding of hospital discharge planning can, then, contribute to our understanding of the health services system. This study involved the development of a process model of discharge planning from hospitals. Model building involved the identification of factors used by discharge planners to develop aftercare plans, and the specification of the roles of these factors in the development of the discharge plan. The factors in the model were concatenated in 16 discrete decision sequences, each of which produced an aftercare plan.^ The sample for this study comprised 407 inpatients admitted to the M. D. Anderson Hospital and Tumor Institution at Houston, Texas, who were discharged to any site within Texas during a 15 day period. Allogeneic bone marrow donors were excluded from the sample. The factors considered in the development of discharge plans were recorded by discharge planners and were used to develop the model. Data analysis consisted of sorting the discharge plans using the plan development factors until for some combination and sequence of factors all patients were discharged to a single site. The arrangement of factors that led to that aftercare plan became a decision sequence in the model.^ The model constructs the same discharge plans as those developed by hospital staff for every patient in the study. Tests of the validity of the model should be extended to other patients at the MDAH, to other cancer hospitals, and to other inpatient services. Revisions of the model based on these tests should be of value in the management of discharge planning services and in the design and development of comprehensive community health services.^

AN ECONOMIC ANALYSIS OF THE DEMAND FOR CARE IN A FREE-STANDING EMERGENCY CENTER

Free-standing emergency centers (FECs) represent a new approach to the delivery of health care which are competing for patients with more conventional forms of ambulatory care in many parts of the U.S. Currently, little is known about these centers and their patient populations. The purpose of this study, therefore, was to describe the patients who visited two commonly-owned FECs, and determine the reasons for their visits. An economic model of the demand for FEC care was developed to test its ability to predict the economic and sociodemographic factors of use. Demand analysis of other forms of ambulatory services, such as a regular source of care (RSOC), was also conducted to examine the issues of substitution and complementarity.^ A systematic random sample was chosen from all private patients who used the clinics between July 1 and December 31, 1981. Data were obtained by means of a telephone interview and from clinic records. Five hundred fifty-one patients participated in the study.^ The typical FEC patient was a 26 year old white male with a minimum of a high school education, and a family income exceeding $25,000 a year. He had lived in the area for at least twenty years, and was a professional or a clerical worker. The patients made an average of 1.26 visits to the FECs in 1981. The majority of the visits involved a medical complaint; injuries and preventive care were the next most common reasons for visits.^ The analytic results revealed that time played a relatively important role in the demand for FEC care. As waiting time at the patients' regular source of care increased, the demand for FEC care increased, indicating that the clinic serves as a substitute for the patients' usual means of care. Age and education were inversely related to the demand for FEC care, while those with a RSOC frequented the clinics less than those lacking such a source.^ The patients used the familiar forms of ambulatory care, such as a private physician or an emergency room in a more typical fashion. These visits were directly related to the age and education of the patients, existence of a regular source of care, and disability days, which is a measure of health status. ^

TIME SERIES ANALYSIS AS INPUT FOR PREDICTIVE MODELING: PREDICTING CARDIAC ARREST IN A PEDIATRIC INTENSIVE CARE UNIT

The first manuscript, entitled "Time-Series Analysis as Input for Clinical Predictive Modeling: Modeling Cardiac Arrest in a Pediatric ICU" lays out the theoretical background for the project. There are several core concepts presented in this paper. First, traditional multivariate models (where each variable is represented by only one value) provide single point-in-time snapshots of patient status: they are incapable of characterizing deterioration. Since deterioration is consistently identified as a precursor to cardiac arrests, we maintain that the traditional multivariate paradigm is insufficient for predicting arrests. We identify time series analysis as a method capable of characterizing deterioration in an objective, mathematical fashion, and describe how to build a general foundation for predictive modeling using time series analysis results as latent variables. Building a solid foundation for any given modeling task involves addressing a number of issues during the design phase. These include selecting the proper candidate features on which to base the model, and selecting the most appropriate tool to measure them. We also identified several unique design issues that are introduced when time series data elements are added to the set of candidate features. One such issue is in defining the duration and resolution of time series elements required to sufficiently characterize the time series phenomena being considered as candidate features for the predictive model. Once the duration and resolution are established, there must also be explicit mathematical or statistical operations that produce the time series analysis result to be used as a latent candidate feature. In synthesizing the comprehensive framework for building a predictive model based on time series data elements, we identified at least four classes of data that can be used in the model design. The first two classes are shared with traditional multivariate models: multivariate data and clinical latent features. Multivariate data is represented by the standard one value per variable paradigm and is widely employed in a host of clinical models and tools. These are often represented by a number present in a given cell of a table. Clinical latent features derived, rather than directly measured, data elements that more accurately represent a particular clinical phenomenon than any of the directly measured data elements in isolation. The second two classes are unique to the time series data elements. The first of these is the raw data elements. These are represented by multiple values per variable, and constitute the measured observations that are typically available to end users when they review time series data. These are often represented as dots on a graph. The final class of data results from performing time series analysis. This class of data represents the fundamental concept on which our hypothesis is based. The specific statistical or mathematical operations are up to the modeler to determine, but we generally recommend that a variety of analyses be performed in order to maximize the likelihood that a representation of the time series data elements is produced that is able to distinguish between two or more classes of outcomes. The second manuscript, entitled "Building Clinical Prediction Models Using Time Series Data: Modeling Cardiac Arrest in a Pediatric ICU" provides a detailed description, start to finish, of the methods required to prepare the data, build, and validate a predictive model that uses the time series data elements determined in the first paper. One of the fundamental tenets of the second paper is that manual implementations of time series based models are unfeasible due to the relatively large number of data elements and the complexity of preprocessing that must occur before data can be presented to the model. Each of the seventeen steps is analyzed from the perspective of how it may be automated, when necessary. We identify the general objectives and available strategies of each of the steps, and we present our rationale for choosing a specific strategy for each step in the case of predicting cardiac arrest in a pediatric intensive care unit. Another issue brought to light by the second paper is that the individual steps required to use time series data for predictive modeling are more numerous and more complex than those used for modeling with traditional multivariate data. Even after complexities attributable to the design phase (addressed in our first paper) have been accounted for, the management and manipulation of the time series elements (the preprocessing steps in particular) are issues that are not present in a traditional multivariate modeling paradigm. In our methods, we present the issues that arise from the time series data elements: defining a reference time; imputing and reducing time series data in order to conform to a predefined structure that was specified during the design phase; and normalizing variable families rather than individual variable instances. The final manuscript, entitled: "Using Time-Series Analysis to Predict Cardiac Arrest in a Pediatric Intensive Care Unit" presents the results that were obtained by applying the theoretical construct and its associated methods (detailed in the first two papers) to the case of cardiac arrest prediction in a pediatric intensive care unit. Our results showed that utilizing the trend analysis from the time series data elements reduced the number of classification errors by 73%. The area under the Receiver Operating Characteristic curve increased from a baseline of 87% to 98% by including the trend analysis. In addition to the performance measures, we were also able to demonstrate that adding raw time series data elements without their associated trend analyses improved classification accuracy as compared to the baseline multivariate model, but diminished classification accuracy as compared to when just the trend analysis features were added (ie, without adding the raw time series data elements). We believe this phenomenon was largely attributable to overfitting, which is known to increase as the ratio of candidate features to class examples rises. Furthermore, although we employed several feature reduction strategies to counteract the overfitting problem, they failed to improve the performance beyond that which was achieved by exclusion of the raw time series elements. Finally, our data demonstrated that pulse oximetry and systolic blood pressure readings tend to start diminishing about 10-20 minutes before an arrest, whereas heart rates tend to diminish rapidly less than 5 minutes before an arrest.